|
(11) | EP 4 105 770 A1 |
| (12) | EUROPEAN PATENT APPLICATION |
| published in accordance with Art. 153(4) EPC |
|
|
|
|
|||||||||||||||||||||||||||
| (54) | B+ TREE ACCESS METHOD AND APPARATUS, AND COMPUTER-READABLE STORAGE MEDIUM |
| (57) Provided are a B+ tree access method and apparatus, and a medium. The method includes:
when a B+ tree creation instruction is acquired, determining whether there is an idle
underlying tree file mapped to a Dynamic Random Access Memory (DRAM); if YES, storing
underlying data of a B+ tree in the underlying tree file; if NO, creating a new target
underlying tree file, and mapping the target underlying tree file to the DRAM, so
as to store the underlying data of the B+ tree in the target underlying tree file;
and when a number of layers of the B+ tree is greater than or equal to a preset threshold,
storing, in a preset storage area, data of a layer numbered to be greater than or
equal to the preset threshold in the B+ tree. On the basis of a data structure of
the B+ tree, data needs to be accessed from the underlying data whenever read. An
underlying file is stored in the idle underlying tree file mapped to the DRAM, whereby
the efficiency of accessing the underlying data is effectively improved. The other
data is stored in a storage space other than the DRAM, whereby the utilization rate
of DRAM resources is increased. |
[0001] This application claims priority to Chinese Patent Application No. 202010093601.7, filed on February 14, 2020, in China National Intellectual Property Administration and entitled "B+ Tree Access Method and Apparatus, and Computer-Readable Storage Medium", the contents of which are hereby incorporated by reference in its entirety.
FIELD
[0002] The present disclosure relates to the technical field of data storage, and particularly to a B+ tree access method and apparatus, and a computer-readable storage medium.
BACKGROUND
[0003] A plenty of B+ tree structures are used in an all-flash array storage system because of a special structure thereof. For example, an all-flash array generally uses thin provisioning, and a logical address of a volume is in no longer linear correspondence but approximately random mapping with a physical address of the volume on a Redundant Array of Independent Disks (RAID). In order to manage such mapping relationships, B+ trees are used to store mappings from logical addresses to physical addresses of volumes and inverse mappings from the physical address to the logical address. For a de-duplication function of the all-flash array, B+ trees are used to store mappings from HASH values of data blocks to physical addresses.
[0004] A Dynamic Random Access Memory (DRAM) is a relatively common system memory. Since the DRAM is a relatively expensive component, only a relatively small number of DRAMs are configured in a storage system to reduce the cost. In the prior art, data of a B+ tree is usually stored in a Solid State Drive (SSD). Data in the SSD needs to be read into the memory first and then read/written, while an Input/Output (10) path for reading/writing the SSD is relatively long, so the speed is low. Frequently reading a B+ tree structure into the memory and swapping the B+ tree structure out of the memory may bring a high Central Processing Unit (CPU) overhead.
[0005] Therefore, how to improve the B+ tree read/write efficiency is a problem that needs to be solved by those skilled in the art.
SUMMARY
[0006] An objective of embodiments of the present disclosure is to provide a B+ tree access method and apparatus, and a computer-readable storage medium, thereby improving the B+ tree read/write efficiency.
[0007] In order to solve the foregoing technical problem, the embodiments of the present disclosure provide a B+ tree access method, including:
when a B+ tree creation instruction is acquired, determining whether there is an idle underlying tree file mapped to a DRAM;
if YES, storing underlying data of a B+ tree in the underlying tree file;
if NO, creating a new target underlying tree file, and mapping the target underlying tree file to the DRAM, so as to store the underlying data of the B+ tree in the target underlying tree file; and
when a number of layers of the B+ tree is greater than or equal to a preset threshold, storing, in a preset storage area, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree.
[0008] In an embodiment, the B+ tree correspondingly includes a first layer, a second layer, and a third layer, and data of the first layer is determined as the underlying data;
[0009] Correspondingly, when the number of layers of the B+ tree is greater than or equal to the preset threshold, storing, in a preset storage area, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree includes:
storing data of the second layer of the B+ tree in a Data Center Persistent Memory Module (DCPMM); and
storing data of the third layer of the B+ tree in a preset hard disk.
[0010] In a further embodiment, before the B+ tree creation instruction is acquired, determining whether there is an idle underlying tree file mapped to a DRAM, the method further includes:
determining a minimum read/write granularity of the DCPMM as a node capacity of the B+ tree.
[0011] In a further embodiment, the step of storing underlying data of the B+ tree in the underlying tree file includes:
storing the underlying data of the B+ tree in the underlying tree file as a separate piece of node data according to the node capacity, wherein an offset address of a node of a next layer is stored in a key-value pair of the separate piece of node data.
[0012] In a further embodiment, the method further includes:
when a data query instruction is acquired, determining a root node according to a logical address contained in the data query instruction; and
determining a leaf node according to an offset address in the root node, and reading data corresponding to the leaf node.
[0013] In a further embodiment, the method further includes:
when a data modification instruction is acquired, determining whether there is node data matched with a node identifier contained in the data modification instruction in the DRAM;
if YES, modifying the node data according to surrogate data contained in the data modification instruction, and setting a dirty flag for modified node data;
if NO, determining whether there is the node data matched with the node identifier contained in the data modification instruction in the DCPMM;
when there is the node data matched with the node identifier contained in the data modification instruction in the DCPMM, modifying the node data according to surrogate data contained in the data modification instruction;
when there is no node data matched with the node identifier contained in the data modification instruction in the DCPMM, reading the node data matched with the node identifier contained in the data modification instruction from the hard disk to the DRAM; and completing, in the DRAM, modification of the node data according to surrogate data contained in the data modification instruction, and setting the dirty flag for modified node data.
[0014] In a further embodiment, after storing underlying data of the B+ tree in the underlying tree file, the method further includes:
migrating data mapped to the DRAM and set with the dirty flag to the DCPMM according to preset periodic time.
[0015] The embodiments of the present disclosure also provide a B+ tree access apparatus, including a first judgment unit, a first storage unit, a creation unit, and a second storage unit.
[0016] The first judgment unit is configured to, when a B+ tree creation instruction is acquired, determine whether there is an idle underlying tree file mapped to a DRAM, if YES, trigger the first storage unit, or if NO, trigger the creation unit.
[0017] The first storage unit is configured to store underlying data of a B+ tree in the underlying tree file.
[0018] The creation unit is configured to create a new target underlying tree file, and map the target underlying tree file to the DRAM, so as to store the underlying data of the B+ tree in the target underlying tree file.
[0019] The second storage unit is configured to, when a number of layers of the B+ tree is greater than or equal to a preset threshold, store, in a preset storage area, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree.
[0020] In an embodiment, the B+ tree correspondingly includes a first layer, a second layer, and a third layer, and data of the first layer is determined as the underlying data.
[0021] Correspondingly, the second storage unit is specifically configured to store data of the second layer of the B+ tree in a DCPMM, and store data of the third layer of the B+ tree in a preset hard disk.
[0023] The determining-as unit is configured to determine a minimum read/write granularity of the DCPMM as a node capacity of the B+ tree.
[0024] In a further embodiment, the first storage unit is specifically configured to store the underlying data of the B+ tree in the underlying tree file as a separate piece of node data according to the node capacity, wherein an offset address of a node of a next layer is stored in a key-value pair of the separate piece of node data.
[0025] In a further embodiment, the apparatus further includes a query unit, a determination unit, and a reading unit.
[0026] The query unit is configured to, when a data query instruction is acquired, determine a root node according to a logical address contained in the data query instruction.
[0027] The determination unit is configured to determine a leaf node according to an offset address in the root node.
[0029] In a further embodiment, the apparatus further includes a second judgment unit, a modification unit, a setting unit, a third judgment unit, and a reading unit.
[0030] The second judgment unit is configured to, when a data modification instruction is acquired, determine whether there is node data matched with a node identifier contained in the data modification instruction in the DRAM, if YES, trigger the modification unit, or if NO, trigger the third judgment unit.
[0031] The modification unit is configured to modify the node data according to surrogate data contained in the data modification instruction.
[0033] The third judgment unit determines whether there is the node data matched with the node identifier contained in the data modification instruction in the DCPMM
[0034] The modification unit is further configured to, when there is the node data matched with the node identifier contained in the data modification instruction in the DCPMM, modify the node data according to surrogate data contained in the data modification instruction.
[0035] The reading unit is configured to, when there is no node data matched with the node identifier contained in the data modification instruction in the DCPMM, read the node data matched with the node identifier contained in the data modification instruction from the hard disk to the DRAM
[0036] The modification unit is further configured to complete, in the DRAM, modification of the node data according to surrogate data contained in the data modification instruction, and trigger the setting unit to set the dirty flag for modified node data.
[0038] The migration unit is configured to migrate the data mapped to the DRAM and set with the dirty flag to the DCPMM according to preset periodic time.
[0039] The embodiments of the present disclosure also provide a B+ tree access apparatus, including:
a memory, configured to store a computer program; and
a processor, configured to execute the computer program to implement any steps of the B+ tree access method as described above.
[0040] The embodiments of the present disclosure also provide a computer-readable storage medium, storing a computer program that is executed by a processor to implement any steps of the B+ tree access method as described above.
[0041] It can be seen from the above technical solutions that, when a B+ tree creation instruction is acquired, whether there is an idle underlying tree file mapped to a DRAM is determined. When there is the idle underlying tree file mapped to the DRAM, underlying data of a B+ tree may be directly stored in the underlying tree file. When there is no idle underlying tree file mapped to the DRAM, it is necessary to create a new target underlying tree file and map the target underlying tree file to the DRAM, so as to store the underlying data of the B+ tree in the target underlying tree file. When a number of layers of the B+ tree is greater than or equal to a preset threshold, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree is stored in a preset storage area. On the basis of a data structure of the B+ tree, data needs to be accessed from the underlying data whenever read, so the underlying data is accessed more. An underlying file is stored in the idle underlying tree file mapped to the DRAM, whereby the efficiency of accessing the underlying data is effectively improved. The other data in the B+ tree except the underlying data is accessed less, and in order to reduce the occupation of the DRAM, the other data may be stored in a storage space other than the DRAM. Therefore, the utilization rate of DRAM resources is increased, and the B+ tree read/write efficiency may also be ensured.
BRIEF DESCRIPTION OF THE DRAWINGS
[0042] In order to describe the embodiments of the present disclosure more clearly, the drawings required to be used in the embodiments will be introduced briefly below. Apparently, the drawings in the description below are only some embodiments of the present disclosure. Those ordinarily skilled in the art may further obtain other drawings according to these drawings without creative work.
Fig. 1 is a flowchart of a B+ tree access method according to an embodiment of the present disclosure;
Fig. 2 is a flowchart of a B+ tree data modification method according to an embodiment of the present disclosure;
Fig. 3 is a schematic structural diagram of a B+ tree access apparatus according to an embodiment of the present disclosure; and
Fig. 4 is a schematic diagram of a hardware structure of a B+ tree access apparatus according to an embodiment of the present disclosure.
DETAILED DESCRIPTION
[0043] The technical solutions in the embodiments of the present disclosure will be described clearly and completely below in combination with the drawings in the embodiments of the present disclosure. Clearly, the described embodiments are not all but only part of embodiments of the present disclosure. All other embodiments obtained by those ordinarily skilled in the art based on the embodiments in the present disclosure without creative work shall fall within the scope of protection of the present disclosure.
[0044] In order to make the solutions of the present disclosure understood better by those skilled in the art, the present disclosure will be further described below in detail in combination with the drawings and specific embodiments.
[0045] AB+ tree access method provided in the embodiments of the present disclosure is then introduced in detail. Fig. 1 is a flowchart of a B+ tree access method according to an embodiment of the present disclosure. The method includes the following steps.
[0046] In S101, when a B+ tree creation instruction is acquired, whether there is an idle underlying tree file mapped to a DRAM is determined.
[0047] When there is the idle underlying tree file mapped to the DRAM, it indicates that there is a file available for storing a B+ tree in the DRAM. In such case, S102 may be performed.
[0048] When there is no idle underlying tree file mapped to the DRAM, it indicates that there is no file for storing a B+ tree in the current DRAM. In such case, S103 may be performed.
[0050] In S103, a new target underlying tree file is created, and the target underlying tree file is mapped to the DRAM, so as to store the underlying data of the B+ tree in the target underlying tree file.
[0051] On the basis of a data structure of the B+ tree, data needs to be accessed from the underlying data whenever read, so the underlying data is accessed more frequently. An underlying file is stored in the idle underlying tree file mapped to the DRAM, whereby the efficiency of accessing the underlying data is effectively improved.
[0052] It is to be noted that, for ease of distinction, in the embodiment of the present disclosure, a newly created underlying tree file is referred to as a target underlying tree file. For a storage system, both the newly created target underlying tree file and the underlying tree file are files for storing the underlying data, and there are no substantial differences therebetween.
[0053] In S104, when a number of layers of the B+ tree is greater than or equal to a preset threshold, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree is stored in a preset storage area.
[0054] The B+ tree includes multiple layers of data. In the embodiment of the present disclosure, the data of the B+ tree may be divided into different layers, and frequencies and numbers of times of accessing the data of different layers are different. Therefore, the data of different layers may be stored at different positions.
[0055] In a specific implementation, the B+ tree may be divided into a first layer, a second layer, and a third layer. The data of the first layer is determined as the underlying data.
[0056] At the beginning of creation of the B+ tree, the underlying data is accessed more frequently. Therefore, in the embodiment of the present disclosure, the underlying data is stored in the idle underlying tree file mapped to the DRAM
[0057] Considering that a DCPMM may implement direct data access without the DRAM in a Device Direct Access (DAX) mode, a DCPMM device in the Device DAX mode may be regarded as a DCPMM. In the embodiment of the present disclosure, data of the second layer of the B+ tree may be stored in a DCPMM
[0058] In order to accurately control the use of the DCPMM and avoid DCPMM resources being wasted, data of the third layer of the B+ tree may be stored in a preset hard disk.
[0059] In practical applications, a first-layer file to a third-layer file of the B+ tree may be determined as the first layer, a fourth-layer file and a fifth-layer file may be determined as the second layer, and the data above a fifth-layer file may be determined as the third layer.
[0060] It can be seen from the above technical solutions that, when a B+ tree creation instruction is acquired, whether there is an idle underlying tree file mapped to a DRAM is determined. When there is the idle underlying tree file mapped to the DRAM, underlying data of a B+ tree may be directly stored in the underlying tree file. When there is no idle underlying tree file mapped to the DRAM, it is necessary to create a new target underlying tree file and map the target underlying tree file to the DRAM, so as to store the underlying data of the B+ tree in the target underlying tree file. When a number of layers of the B+ tree is greater than or equal to a preset threshold, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree is stored in a preset storage area. On the basis of a data structure of the B+ tree, data needs to be accessed from the underlying data whenever read, so the underlying data is accessed more. An underlying file is stored in the idle underlying tree file mapped to the DRAM, whereby the efficiency of accessing the underlying data is effectively improved. The other data in the B+ tree except the underlying data is accessed less, and in order to reduce the occupation of the DRAM, the other data may be stored in a storage space other than the DRAM. Therefore, the utilization rate of DRAM resources is increased, and the B+ tree read/write efficiency may also be ensured.
[0061] At the beginning of creation of the B+ tree, the underlying data is accessed more frequently, and thus may be stored in the DRAM. As storage time increases, the underlying data of the B+ tree is no longer accessed so frequently. In order to increase the utilization rate of the DRAM, during specific implementation, a dirty flag may be set for the underlying data stored in the DRAM, and then the underlying data stored in the DRAM may be migrated to the DCPMM periodically.
[0062] In order to improve the data read efficiency, during specific implementation, a minimum read/write granularity of the DCPMM may be determined as a node capacity of the B+ tree. An internal minimum read/write granularity of the DCPMM is 256 bytes, so a capacity of a B+ tree node may be set to 256 bytes, namely a set of check protection data is generated by every 256 bytes of data.
[0063] In the embodiment of the present disclosure, the underlying data of the B+ tree may be stored in the underlying tree file as a separate piece of node data according to the node capacity. An offset address of a node of a next layer is stored in a key-value pair of the separate piece of node data.
[0064] In practical applications, each node may include a 16-byte node header and at most 15 key-value pairs. The node header includes a node type and other information. The node type includes a leaf node and a non-leaf node. For a leaf node, a value in a key-value pair thereof is a RAID physical address, and specific data corresponding to the leaf node may be read according to the RAID physical address. For a non-leaf node, an offset address of a node of a next layer may be stored in a key-value pair thereof.
[0065] When a data query instruction is acquired, a root node may be determined according to a logical address contained in the data query instruction. A leaf node is determined according to an offset address in the root node. Data corresponding to the leaf node is read.
[0066] For example, when data of the B+ tree needs to be queried, a logical address of a data volume may be contained in a data query instruction. Metadata of the data volume may be found according to the logical address. If a memory address of a root node is protected in the metadata, an underlying tree file of the tree is in the memory. A leaf node may be searched according to an offset address stored in the corresponding node in the underlying tree file. If no leaf node is found till a third-layer node, an offset address of the corresponding node in a third-layer file is read from the metadata, thereby obtaining a storage address of a fourth-layer file. If this node is still a non-leaf node, a fifth-layer node continues to be accessed. If the fifth-layer node is still not a leaf node, the third-layer file is read into the memory to continue searching until a leaf node is found. Specific data corresponding to the leaf node may be read according to a RAID physical address stored in a key-value pair of the leaf node.
[0067] The data of the B+ tree is stored at different positions according to the divided layers. When data of the B+ tree needs to be modified, the storage position of the data that needs to be modified may be queried according to the storage positions, thereby implementing modification of the data in the B+ tree. Fig. 2 is a flowchart of a B+ tree data modification method according to an embodiment of the present disclosure, including the following steps.
[0068] In S201, when a data modification instruction is acquired, whether there is node data matched with a node identifier contained in the data modification instruction in the DRAM is determined.
[0069] When there is the node data matched with the node identifier contained in the data modification instruction in the DRAM, S202 is performed.
[0070] When there is no node data matched with the node identifier contained in the data modification instruction in the DRAM, it is necessary to further determine a position of the node data that needs to be modified. In such case, S203 may be performed.
[0071] In S202, the node data is modified according to surrogate data contained in the data modification instruction, and a dirty flag is set for modified node data.
[0072] Storing the node data in the DRAM may occupy DRAM resources. In order to increase the utilization rate of the DRAM resources, a dirty flag may be set for modified node data, and then the data mapped to the DRAM and set with the dirty flag may subsequently be migrated to the DCPMM according to preset periodic time.
[0073] In S203, whether there is the node data matched with the node identifier contained in the data modification instruction in the DCPMM is determined.
[0074] When there is the node data matched with the node identifier contained in the data modification instruction in the DCPMM, S204 is performed.
[0075] In the embodiment of the present disclosure, the data of the B+ tree may be stored in the DRAM, the DCPMM, and the preset hard disk respectively according to different layers. When there is no node data matched with the node identifier contained in the data modification instruction in the DRAM, and there is no node data matched with the node identifier contained in the data modification instruction in the DCPMM, it indicates that the node data that needs to be modified is stored in the preset hard disk. In such case, S205 may be performed.
[0076] In S204, the node data is modified according to surrogate data contained in the data modification instruction.
[0077] In S205, the node data matched with the node identifier contained in the data modification instruction is read from the hard disk to the DRAM
[0078] Since data in the hard disk is not directly modifiable, when the node data that needs to be modified is stored in the preset hard disk, the node data matched with the node identifier contained in the data modification instruction needs to be read from the hard disk to the DRAM
[0079] In S206, modification of the node data is completed in the DRAM according to surrogate data contained in the data modification instruction, and a dirty flag is set for modified node data.
[0080] When the node data that needs to be modified is stored in the DRAM or the hard disk, the modified node data may be stored in the DRAM. Since there are limited DRAM resources, to avoid the modified node data occupying the DRAM resources for long, in the embodiment of the present disclosure, a dirty flag may be set for the modified node data stored in the DRAM, and then the data mapped to the DRAM and set with the dirty flag may subsequently be migrated to the DCPMM according to preset periodic time.
[0081] The data stored in the DRAM is migrated periodically, whereby the long-term occupation of the DRAM resources by the data of the B+ tree is avoided, and the processing performance of the storage system is improved effectively.
[0082] Fig. 3 is a schematic structural diagram of a B+ tree access apparatus according to an embodiment of the present disclosure, including a first judgment unit 31, a first storage unit 32, a creation unit 33, and a second storage unit 34.
[0083] The first judgment unit 31 is configured to, when a B+ tree creation instruction is acquired, determine whether there is an idle underlying tree file mapped to a DRAM, if YES, trigger the first storage unit 32, or if NO, trigger the creation unit 33.
[0084] The first storage unit 32 is configured to store underlying data of a B+ tree in the underlying tree file.
[0085] The creation unit 33 is configured to create a new target underlying tree file, and map the target underlying tree file to the DRAM, so as to store the underlying data of the B+ tree in the target underlying tree file.
[0086] The second storage unit 34 is configured to, when a number of layers of the B+ tree is greater than or equal to a preset threshold, store, in a preset storage area, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree.
[0087] Optionally, the B+ tree correspondingly includes a first layer, a second layer, and a third layer, and data of the first layer is determined as the underlying data.
[0088] Correspondingly, the second storage unit is specifically configured to store data of the second layer of the B+ tree in a DCPMM, and store data of the third layer of the B+ tree in a preset hard disk.
[0090] The determining-as unit is configured to determine a minimum read/write granularity of the DCPMM as a node capacity of the B+ tree.
[0091] Optionally, the first storage unit is specifically configured to store the underlying data of the B+ tree in the underlying tree file as a separate piece of node data according to the node capacity, wherein an offset address of a node of a next layer is stored in a key-value pair of the separate piece of node data.
[0092] Optionally, the apparatus further includes a query unit, a determination unit, and a reading unit.
[0093] The query unit is configured to, when a data query instruction is acquired, determine a root node according to a logical address contained in the data query instruction.
[0094] The determination unit is configured to determine a leaf node according to an offset address in the root node.
[0096] Optionally, the apparatus further includes a second judgment unit, a modification unit, a setting unit, a third judgment unit, and a reading unit.
[0097] The second judgment unit is configured to, when a data modification instruction is acquired, determine whether there is node data matched with a node identifier contained in the data modification instruction in the DRAM, if YES, trigger the modification unit, or if NO, trigger the third judgment unit.
[0098] The modification unit is configured to modify the node data according to surrogate data contained in the data modification instruction.
[0100] The third judgment unit determines whether there is the node data matched with the node identifier contained in the data modification instruction in the DCPMM
[0101] The modification unit is further configured to, when there is the node data matched with the node identifier contained in the data modification instruction in the DCPMM, modify the node data according to surrogate data contained in the data modification instruction.
[0102] The reading unit is configured to, when there is no node data matched with the node identifier contained in the data modification instruction in the DCPMM, read the node data matched with the node identifier contained in the data modification instruction from the hard disk to the DRAM
[0103] The modification unit is further configured to complete, in the DRAM, modification of the node data according to surrogate data contained in the data modification instruction, and trigger the setting unit to set a dirty flag for modified node data.
[0105] The migration unit is configured to migrate the data mapped to the DRAM and set with the dirty flag to the DCPMM according to preset periodic time.
[0106] Descriptions about features in the embodiment corresponding to Fig. 3 may refer to the related descriptions in the embodiments corresponding to Figs. 1 and 2. Elaborations are omitted herein.
[0107] It can be seen from the above technical solutions that, when a B+ tree creation instruction is acquired, whether there is an idle underlying tree file mapped to a DRAM is determined. When there is the idle underlying tree file mapped to the DRAM, underlying data of a B+ tree may be directly stored in the underlying tree file. When there is no idle underlying tree file mapped to the DRAM, it is necessary to create a new target underlying tree file and map the target underlying tree file to the DRAM, so as to store the underlying data of the B+ tree in the target underlying tree file. When a number of layers of the B+ tree is greater than or equal to a preset threshold, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree is stored in a preset storage area. On the basis of a data structure of the B+ tree, data needs to be accessed from the underlying data whenever read, so the underlying data is accessed more. An underlying file is stored in the idle underlying tree file mapped to the DRAM, whereby the efficiency of accessing the underlying data is effectively improved. The other data in the B+ tree except the underlying data is accessed less, and in order to reduce the occupation of the DRAM, the other data may be stored in a storage space other than the DRAM. Therefore, the utilization rate of DRAM resources is increased, and the B+ tree read/write efficiency may also be ensured.
[0108] Fig. 4 is a schematic diagram of a hardware structure of a B+ tree access apparatus 40 according to an embodiment of the present disclosure, including:
a memory 41, configured to store a computer program; and
a processor 42, configured to execute the computer program to implement any steps of the B+ tree access method as described in any one of the above-mentioned embodiments.
[0109] The embodiments of the present disclosure also provide a computer-readable storage medium, storing a computer program that is executed by a processor to implement any steps of the B+ tree access method as described in any one of the above-mentioned embodiments.
[0110] The B+ tree access method and apparatus and computer-readable storage medium provided in the embodiments of the present disclosure are introduced above in detail. Each embodiment in the specification is described progressively. Descriptions made in each embodiment focus on differences from the other embodiments, and the same or similar parts in each embodiment refer to the other embodiments. The apparatus disclosed in the embodiments corresponds to the method disclosed in the embodiments, and thus is described relatively briefly, and related parts refer to the descriptions about the method. It is to be pointed out that those ordinarily skilled in the art may further make a plurality of improvements and embellishments to the present disclosure without departing from the principle of the present disclosure, and these improvements and embellishments shall also fall within the scope of protection of the claims of the present disclosure.
[0111] Professionals may further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein may be implemented by electronic hardware, computer software, or a combination thereof. In order to describe the interchangeability of hardware and software clearly, the compositions and steps of each example have been described generally in the foregoing descriptions according to functions. Whether these functions are executed by hardware or software depends on specific applications and design constraints of the technical solutions. Professionals may realize the described functions for each specific application by use of different methods, but such realization shall fall within the scope of the present disclosure.
[0112] The steps of the method or algorithm described in combination with the embodiments disclosed herein may directly be implemented by hardware, a software module executed by the processor, or a combination thereof. The software module may be arranged in a Random Access Memory (RAM), an internal memory, a Read-Only Memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a Compact Disc Read-Only Memory (CD-ROM), or a storage medium in any other form well known in this art.
1. A B+ tree access method, characterized by comprising:
when a B+ tree creation instruction is acquired, determining whether there is an idle underlying tree file mapped to a Dynamic Random Access Memory (DRAM);
if YES, storing underlying data of a B+ tree in the underlying tree file;
if NO, creating a new target underlying tree file, and mapping the target underlying tree file to the DRAM, so as to store the underlying data of the B+ tree in the target underlying tree file; and
when a number of layers of the B+ tree is greater than or equal to a preset threshold, storing, in a preset storage area, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree.
2. The method according to claim 1, characterized in that the B+ tree correspondingly comprises a first layer, a second layer, and a third
layer, and data of the first layer is determined as the underlying data; and
correspondingly, when the number of layers of the B+ tree is greater than or equal to the preset threshold, storing, in a preset storage area, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree comprises:
correspondingly, when the number of layers of the B+ tree is greater than or equal to the preset threshold, storing, in a preset storage area, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree comprises:
storing data of the second layer of the B+ tree in a Data Center Persistent Memory Module (DCPMM), and
storing data of the third layer of the B+ tree in a preset hard disk.
3. The method according to claim 2, characterized in that before the B+ tree creation instruction is acquired, determining whether there is
an idle underlying tree file mapped to a DRAM, the method further comprises:
determining a minimum read/write granularity of the DCPMM as a node capacity of the B+ tree.
determining a minimum read/write granularity of the DCPMM as a node capacity of the B+ tree.
4. The method according to claim 3, characterized in that the step of storing underlying data of the B+ tree in the underlying tree file comprises:
storing the underlying data of the B+ tree in the underlying tree file as a separate piece of node data according to the node capacity, wherein an offset address of a node of a next layer is stored in a key-value pair of the separate piece of node data.
storing the underlying data of the B+ tree in the underlying tree file as a separate piece of node data according to the node capacity, wherein an offset address of a node of a next layer is stored in a key-value pair of the separate piece of node data.
5. The method according to claim 4, characterized by further comprising:
when a data query instruction is acquired, determining a root node according to a logical address contained in the data query instruction; and
determining a leaf node according to an offset address in the root node, and reading data corresponding to the leaf node.
6. The method according to claim 2, characterized by further comprising:
when a data modification instruction is acquired, determining whether there is node data matched with a node identifier contained in the data modification instruction in the DRAM;
if YES, modifying the node data according to surrogate data contained in the data modification instruction, and setting a dirty flag for modified node data;
if NO, determining whether there is the node data matched with the node identifier contained in the data modification instruction in the DCPMM;
when there is the node data matched with the node identifier contained in the data modification instruction in the DCPMM, modifying the node data according to surrogate data contained in the data modification instruction;
when there is no node data matched with the node identifier contained in the data modification instruction in the DCPMM, reading the node data matched with the node identifier contained in the data modification instruction from the hard disk to the DRAM; and completing, in the DRAM, modification of the node data according to surrogate data contained in the data modification instruction, and setting the dirty flag for modified node data.
7. The method according to claim 6, characterized in that after storing underlying data of the B+ tree in the underlying tree file, the method
further comprises:
migrating data mapped to the DRAM and set with the dirty flag to the DCPMM according to preset periodic time.
migrating data mapped to the DRAM and set with the dirty flag to the DCPMM according to preset periodic time.
8. A B+ tree access apparatus, characterized by comprising a first judgment unit, a first storage unit, a creation unit, and a second
storage unit, wherein
the first judgment unit is configured to, when a B+ tree creation instruction is acquired, determine whether there is an idle underlying tree file mapped to a Dynamic Random Access Memory (DRAM), if YES, trigger the first storage unit, or if NO, trigger the creation unit;
the first storage unit is configured to store underlying data of a B+ tree in the underlying tree file;
the creation unit is configured to create a new target underlying tree file, and map the target underlying tree file to the DRAM, so as to store the underlying data of the B+ tree in the target underlying tree file; and
the second storage unit is configured to, when a number of layers of the B+ tree is greater than or equal to a preset threshold, store, in a preset storage area, data of a layer numbered to be greater than or equal to the preset threshold in the B+ tree.
9. A B+ tree access apparatus, characterized by comprising:
a memory, configured to store a computer program; and
a processor, configured to execute the computer program to implement any steps of the B+ tree access method according to any one of claims 1-7.
10. A computer-readable storage medium, characterized by storing a computer program that is executed by a processor to implement any steps
of the B+ tree access method according to any one of claims 1-7.
REFERENCES CITED IN THE DESCRIPTION
This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.
Patent documents cited in the description